ML Market Map — China A-Share Clusters for 2026-10-15

A daily unsupervised machine-learning read of the China A-Share market: 5,484 stocks grouped into 7 consensus clusters (KMeans + Gaussian-mixture + hierarchical, over robust-scaled PCA features) for 2026-10-15. Descriptive, not predictive — there is no buy or sell signal. Research, not investment advice.

Markets: US Equity · China A-Share · Japan Equity · Hong Kong Equity

Today's clusters

low volatility (1y) · high margin · low idio vol — 1,443 names; mostly Industrials; drivers: rv 252 (−0.51), profit margin (+0.50), residual vol 252 d (−0.49)
rising earnings · deep pullbacks (21d) · falling (3m) — 1,131 names; mostly Industrials; drivers: ni growth (+0.56), ret 126d (−0.52), pa mae 21d (−0.49)
low margin · expensive (low E/P) · high leverage — 946 names; mostly Industrials; drivers: profit margin (−1.66), val ep z (−1.64), leverage debt to mcap (+1.00)
wide 1y upside range · strong 1y momentum · favorable 1y edge — 658 names; mostly Technology; drivers: pa mfe 52w (+1.90), ret 252d (+1.80), pa edge 52w (+1.72)
high turnover · attention spent · wide upside excursions (21d) — 550 names; mostly Consumer Cyclical; drivers: turnover to mcap (+2.70), pa sir recovered (+2.29), pa mfe 21d (+2.19)
high leverage · cheap (high B/P) · high cash yield — 468 names; mostly Industrials; drivers: leverage debt to mcap (+3.52), val bp z (+1.54), val cfop z (+1.42)
broad market · slight high leverage — 288 names; mostly Industrials; drivers: cfo growth (+4.04), leverage debt to mcap (+0.31), profit margin (−0.29)

Machine-readable data (free, read-only JSON)

The full map, per-ticker cluster assignments with confidence and anomaly scores, and PCA structure are published as open JSON for automated and AI-analyst consumption:

/api/ml/latest?market=cn — clusters, names, sector mix, representative tickers, top features
/api/ml/tickers?market=cn — per-ticker cluster, confidence, anomaly score, day-over-day migration, tenure
/api/ml/pca?market=cn — PCA components, explained variance, feature loadings

Descriptive market-structure research only. Unsupervised clustering finds structure, not direction; a tight cluster or an anomaly is a starting point for research, never a trade signal.